Matrix Reasoning Examples: How Row and Column Patterns Work (Practice Inside)

A matrix can look complicated simply because several small rules are packed into one grid. Once you see what each example is doing, the puzzle stops being a picture and starts behaving like a system.

Matrix reasoning examples are useful because the same few kinds of relationships appear again and again: features progress, values are distributed across rows or columns, shapes combine, or one feature is removed from another.

The point is not to memorize the exact pictures below. It is to notice the structure behind them. A triangle can be swapped for a circle and the colors can change completely while the underlying logic stays the same.

If you want a step-by-step solving method first, use How to Solve Matrix Reasoning Questions. Here, we will learn mainly by working through examples.

Example 1: A Simple Progression

Start with a pattern where only one feature changes. In this matrix, the number of dots increases by one from left to right. The rows use different colors, but color stays constant within each row.

This is a classic pairwise progression: an attribute changes by a consistent amount between neighboring cells. Size, position, angle, or quantity can all progress in the same way.

In their influential analysis of Raven's Progressive Matrices, Carpenter, Just, and Shell identified quantitative pairwise progression as one of a small set of recurring analytic rules used across matrix items. Their broader point is useful here: very different-looking puzzles can rely on the same underlying relation.

Example 2: Distribution Across the Row

Now remove progression entirely. Suppose every row must contain one circle, one triangle, and one square, and the order shifts one place each time.

Nothing is increasing here. The rule is distribution. Each category value must appear once, and its position moves through the row.

This is where people sometimes invent unnecessary movement rules: “the circle moves two spaces, the triangle moves one…” That description may fit, but the cleaner rule is simply that all three values are distributed across every row and column.

Carpenter and colleagues called this kind of structure a distribution of three. Later research has continued to use this taxonomy because matrix difficulty depends partly on the type and number of rules a solver must coordinate.

Example 3: Two Cells Combine into the Third

The next matrix behaves differently again. Here the first two cells of each row are combined to produce the third. Think of the shapes as transparent layers placed on top of each other.

This is figure addition: visual components from two cells are superimposed to create the third. A related question can use subtraction instead — features shared by two cells disappear, or one cell's elements are removed from another.

Addition and subtraction often feel harder than a simple progression because you have to preserve multiple visual components while comparing cells. That is one reason rule combinations matter. Carpenter and colleagues found that errors increased as items required more rules and more goal management, rather than difficulty depending on one visual feature alone.

Example 4: Two Rules at the Same Time

Harder questions often stop being “a progression problem” or “a distribution problem.” They combine them.

Imagine a matrix in which each row contains one circle, triangle, and square, while the number of internal dots increases from one to three across the columns. The missing cell must satisfy both constraints. If the row needs a triangle but the third column needs three dots, the answer is not merely “triangle” — it is a triangle containing three dots.

This is where matrix reasoning starts to feel demanding. You can identify one rule correctly and still choose the wrong answer because a second feature was left unresolved.

Research on matrix solving supports the importance of strategy here. In a study using Raven's matrices, Gonthier and Thomassin experimentally encouraged an effective solving strategy and found that doing so reduced the relationship between working-memory capacity and performance. In a second experiment, reported strategy use statistically mediated that relationship. If you want to test working-memory updating separately from matrix reasoning, Cognitive Train's N-Back Test targets that neighboring skill. The result is a useful reminder that how you organize the problem matters, especially when several relations must be coordinated.

Example 5: Sometimes the Rule Is What Does Not Change

Not every matrix is built around motion or transformation. Sometimes one attribute stays constant across a row while another attribute changes from row to row.

Suppose every cell in the first row contains a circle, every cell in the second row contains a triangle, and every cell in the third row contains a square. Inside those shapes, the number of dots might vary across columns. If the bottom-right cell is missing, the outer shape is already determined: it must be a square. You then solve the dot pattern separately.

Carpenter, Just, and Shell called this constant in a row. It is easy to miss because the eye is naturally drawn to the parts that change. Matrix reasoning often rewards the opposite question as well: what remains invariant while everything else moves?

This becomes especially important when a constant feature is layered with another rule. The outer shape may stay fixed within a row while orientation progresses across columns, or color may stay fixed while quantity changes. Solve the stable feature first and you have one less thing to keep track of.

Try the Examples as Real Problems

The worked examples above are intentionally transparent. Real matrix questions make the same relationships harder to spot by changing several attributes, adding plausible distractors, or combining rules.

Use the embedded Matrix Reasoning Test below as a check on whether you can recognize the structure without the rule being announced first.

For a full dedicated set, open the Matrix Reasoning Test. If the problem is recognizing the pattern rather than maintaining several rule components at once, the broader Inductive Reasoning Test is another useful comparison.

What the Examples Have in Common

Progression, distribution, and combination look different on the page, but they all ask the same deeper question: what relationship stays consistent as the visible details change?

That is why learning examples is more useful than memorizing pictures. Once you recognize that one item is “distribution of three,” it does not matter whether the values are shapes, colors, positions, line directions, or textures. The surface changes; the relationship survives.

For the larger family of nonverbal puzzles, see Abstract Reasoning Examples and Progressive Matrices. The wider Pattern Recognition section connects matrix problems with sequences, visual rules, memory, and other forms of regularity.

You can also browse Cognitive Train's brain tests collection or explore the broader set of brain training and cognitive training tools from the homepage.